LANEKEEPING WITH SHARED CONTROL

Transcription

1 MDYNAMIX AFFILIATED INSTITUTE OF MUNICH UNIVERSITY OF APPLIED SCIENCES LANEKEEPING WITH SHARED CONTROL WHICH ISSUES HAVE TO BE RESEARCHED? 3rd International Symposium on Advanced Vehicle Technology 1

2 OUTLINE State of the art Why do we need it? Pitfalls! Disadvantages/Dangers it might bring Challenges and potentials Summary and outlook 2

3 STATE OF THE ART [1] Shared Control Situation, where human and machine carry out a task simultaneously. 3

4 STATE OF THE ART Classification Type 1 Type 2 Corrective steering intervention Guidance towards lane center e.g. Ford Focus e.g. BMW 5 series Safety-orientated Comfort-orientated Assisted Torque Lenkunterstützung Lane Boundary Fahrstreifen begrenzung Fahrstreifen Lane mitte Center Lane Boundary Fahrstreifen begrenzung [2] 4

5 STATE OF THE ART Classification Steer-by-Angle and Steer-by-Torque: High lateral guidance performances Better reference- and disturbance reactions obtained with SbA SbA disturbance reaction with higher, but less oscillatory jerk values SbT reference reaction with lower rack assist force amplitudes 5

6 WHY DO WE NEED IT? ADVANTAGES IT BRINGS Pros vs Fully Manual Systems More accurate path tracking Reduced control activity Reduced workload vs Fully Automated Systems Drivers stay in-the-loop Better situational awareness 6

7 PITFALLS! DISADVANTAGES/DANGERS IT MIGHT BRING Cons Disengagement of shared control Misinformation, malfunction Who has authority? Human-machine conflict Overreliance and complacency Rely upon supplementary information Skill loss Ignore visual feedback Over-corrective behavior [3] 7

8 1. Detection of departure intention? Drowsy? Intended? [1] 8

9 1. Detection of departure intention [9] Indicator Algorithms Additional sensors 9

10 2. Interaction between driver and assistant system Requirements on Manual control interface Continuous, direct control Certain manual skills Requirements on Automation interface Discrete input Certain analytical and decision-making skills Requirements on shared control interface? What? Goal and intent Direction and magnitude How? Collaborative Information-rich Negotiation of authority Bidirectional 10

11 2. Interaction between driver and assistant system [1] Sensory channels of communication 11

12 2. Interaction between driver and assistant system Combined? Visual? Fast but distractive Auditory? Vigilant but easily affected Haptic? Direct but confusing 12

13 3. Support strategies MM HH Center Guidance Edge Guidance MM HH Continuous support Support as needed 13

14 3. Support strategies Goal of the system? Relief - Reduce workload and control effort? higher level of automation Extension - Avoid dangers in emergency situations? watch out for the after-effect of assistant systems 14

15 4. Harmony of intervention Requirements on the intervention Adequate correction torque Minimally invasive manner Not too frequent corrections No abrupt/jerky intervention Possible consequences: Counter-corrective action of driver Increased physical/mental effort How to standardize/measure? Subjective assessment? Objective measurement? Steering angle velocity Steering angle acceleration Steering torque gradient Derived Key Performance Indicators (KPIs)? 15

16 5. Ability of adaption How can the system adapt to different Various range of drivers drivers? AI, deep learning, big data... Driver selection? Who should adapt? Assistant system? Driver? 16

17 6. Limit of the assistant system Max. assisted torque: 3 Nm Influencing factors: Steering wheel radius Power assist... New ways to specify the system limits? 17

18 7. Hands off detection Definition of hands on Algorithms Additional sensors 18

19 SUMMARY AND OUTLOOK Goal of LKAS? Watch out for the potential dangers Further research directions? Standardization or more freedom for possibilities? What can we learn from other assistant systems? 19

20 MDYNAMIX AFFILIATED INSTITUTE OF MUNICH UNIVERSITY OF APPLIED SCIENCES Contact Information MdynamiX AG Address: Hessstraße 89, München Tel: Fax: Mail: Web: The rights of use of the images contained in the file belong to MdynamiX AG and may not be used in any other way or made available to third parties!

21 SOURCES [1] Flemisch, Frank et al. (2008), COOPERATIVE CONTROL AND ACTIVE INTERFACES FOR VEHICLE ASSISTANCE AND AUTOMATION, FISITA 2008, F [2] Winner, Hermann et al. (2015), Handbuch Fahrerassistenzsysteme [3] J.C.F. de Winter et al. (2011), Preparing drivers for dangerous situations - A critical reflection on continuous shared control, 2011 IEEE International Conference on Systems, Man, and Cybernetics [9] Junyung Lee et al. (2013), Lane-keeping assistance control algorithm using differential braking to prevent unintended lane departures, Control Engineering Practice 23 (2014)